Liuzhou Steel Runs Driverless Molten Iron Transport With Huawei and China Mobile at Its Fangchenggang Base

Liuzhou Steel Group's Fangchenggang base on the Qisha Peninsula in Guangxi moves molten iron at around 1,500 degrees from its blast furnaces to the converters by train, along rail tracks stretching around 2 kilometers. A team of over 70 shunting staff, drivers and signalmen used to work 8-hour shifts close to the molten metal, relaying instructions by walkie-talkie, operating couplings by hand and transporting the ladles with open tops. To take workers out of these hazardous positions, Liuzhou Steel Group developed an intelligent molten iron transport system with Huawei and China Mobile that integrates the industrial internet, 5G and artificial intelligence. The system has three core components. The first, which the company calls the brain, is an intelligent dispatching system that integrates data from the blast furnaces, the transport lines and the steelmaking plants and autonomously plans tasks and routes for each vehicle carrying molten iron. The second, the torso, is an intelligent locomotive system: driverless locomotives fitted with cameras and radars identify road conditions and obstacles in real time and accelerate, decelerate and brake through onboard decision-making systems. The third, the arm, is an intelligent molten iron vehicle with automatic coupler disconnection and automatic parking, using BeiDou positioning for high-precision automated operation. A former shunter told the Global Times that across the whole operation area there is now basically no one in sight except a few inspection and maintenance personnel. The transport system is part of a wider AI programme at the group. On March 31, 2026, Liuzhou Steel Group, Huawei and China Mobile Guangxi launched the Xuantie Steel Model, which the article describes as Guangxi's first AI large language model for the steel industry. It is built on Huawei's Pangu model and uses a "20+N" scenario system covering ironmaking, steelmaking, rolling, logistics, environmental protection and safety. Deputy general manager Shen Min said the group has deployed 33 AI models in full-process intelligent steelmaking, that an intelligent steelmaking scheduling model has raised production efficiency by 8.5 percent, and that models such as a converter AI steelmaking model and a ladle AI argon-blowing model have lowered the production cost of crude steel by 5 yuan per ton.